Can AI read RFQs from WhatsApp messages and voicemails?
Yes. Quoting.ai Supply takes RFQs from WhatsApp messages and voice notes, voicemails, faxes, email, and PDF attachments, and transcribes voice in English, Spanish, Hebrew, and Yiddish. Every request lands in the same approval inbox as a drafted quote: line items matched against your item file, customer price levels applied from the ERP, and a human confirming before anything sends. That last part answers the real objection. A WhatsApp order is flimsy when it lives in one salesperson's phone; routed into an auditable queue with a written quote sent back, it is a documented transaction like any emailed RFQ.
Which channels can AI read RFQs from?
Quoting.ai Supply reads RFQs the way customers actually send them: email, PDF attachments, WhatsApp texts and voice notes, faxes, and voicemails. Voice comes through in English, Spanish, Hebrew, and Yiddish. Distributors we talk to field Hebrew WhatsApp voice notes in Israel and Yiddish voicemails in the NY metro, and the Yiddish voicemail gets the same treatment as the typed email: extracted line items, matched SKUs, one queue.
The hard part is not hearing the words, it is resolving them. Distributors we talk to see a request as plain as a 10 watt light map to 50+ catalog variants. Matching runs against your own item file, cross-references and aliases included, and any line that cannot be resolved confidently is flagged for the human, not guessed.
What happens to a WhatsApp voice note after it arrives?
It is transcribed and kept with the request. Line items are extracted and matched against your item file, that customer's price levels are pulled from the ERP, and the result lands in the approval inbox as a drafted quote with the original message attached. A reviewer checks the draft against the source and approves or edits it before it goes out.
Approved quotes write back to DDI Inform and Spruce today. Epicor, NetSuite, and SAP are in development, and an ERP-light mode exports clean documents for any other system. Trust is governed by the Human Edit Rate, how often a reviewer has to correct a draft. When it runs low and steady, a desk moves from Assist to Guarded to Autopilot; until then, every quote carries a human sign-off.
Aren't WhatsApp orders flimsy and unconfirmable?
They are, when they live in one salesperson's phone. The thread scrolls away, the confirmation is a thumbs-up, and when the customer disputes a count there is nothing to point at. Distributors we talk to describe one inside salesperson doing about 100 quotes a day and working three hours past close. When orders also arrive in that one person's WhatsApp, they are invisible to the rest of the desk, including whoever covers when he is out.
Routing the channel into a queue fixes the channel's reputation. Every message and voice note arrives with a timestamp, a stored transcript, the extracted line items, the matched SKUs, the applied price levels, and a named approver. The customer gets a written quote back to confirm against. That is more paper trail than most phoned-in orders ever had. The channel was never the problem; the pocket was.
Related questions
Do customers have to change how they send orders?
No. They keep texting the same number and leaving the same voicemails. The change is on your side: every request lands in one approval inbox instead of scattered phones and answering machines.
Which languages do voice notes and voicemails work in?
English, Spanish, Hebrew, and Yiddish. Transcription and line-item extraction run the same way in each, and the request and its transcript stay attached to the drafted quote for review.
Does an approved WhatsApp order actually reach our ERP?
On DDI Inform and Spruce, yes, write-back is live today. Epicor, NetSuite, and SAP integrations are in development, and an ERP-light mode exports clean documents for other systems while native integration is scoped.
See it on your own work
Distributors: two steps and a kickoff call. Estimators: upload a plan on a live trade. Either way, the product proves it or it does not.